ON THE RESPIRATION OF TORPEDO MARMORATA

J. exp. Bol. (978), 73, 85-05 S5 Wth 4 fgures Prnted n Great Brtan ON THE RESPIRATION OF TORPEDO MARMORATA BY G. M. HUGHES Insttut de Bologe marne d'arcachon and Research Unt for Comparatve Anmal Respraton, Unversty of Brstol, Woodland Road, Brstol BS8 UG (Receved 9 June 977) SUMMARY. The oxygen consumpton of restng Torpedo marmorata was measured usng three dfferent methods. The results ndcate that ths speces has a much lower oxygen consumpton than other elasmobranchs of comparable sze. 2. The glls are ventlated by a mechansm smlar to that of other rays, but a relatvely small spracular openng seems to be assocated wth a more mportant role of the oro-branchal pump. Durng hypoxa there s a marked ncrease n both frequency and ampltude of the ventlatory movements. 3. The frequency of the heart beat s low and shows lttle change durng hypoxa, except under extreme condtons when bradycarda occurs. 4. In some ndvduals, couplng between cardac and ventlatory pumps s relatvely low but seems to ncrease at lower ventlatory frequences and when the rato between the ventlatory and cardac frequences s a whole number. 5. Extreme hypoxa can be wthstood for many hours but eventually the ventlatory rhythm ceases; t does not recommence mmedately followng a rse n ambent oxygen tenson. 6. The blood has a low oxygen-carryng capacty and a hgh affnty. 7. The surface area of the glls s smaller than that of other speces that have been nvestgated, but the quantty of oxygen transferred/unt surface area s smlar to that known for other speces. 8. It s concluded that Torpedo s a sluggsh fsh adapted to condtons of low oxygen, but the condtons under whch ths occurs reman to be determned. INTRODUCTION One problem met durng nvestgatons of fsh respraton concerns ther actvty and ts varaton durng experments. Many workers have attempted to quantfy such actvty, and some ngenous apparatus has been developed n recent years, whch enables studes to be made of fsh when swmmng at controlled speeds (Brett, 964, 972; Webb, 97; Beamsh, 970). Another soluton has been to study fsh whch spend a substantal amount of tme n an nactve condton; among teleosts, flatfsh such as place and other bottom-lvng forms such as Callonymus (Hughes & Umezawa, 9686; Hughes & Knghts, 968) have provded nformaton of general nterest. Among elasmobranchs, bottom-lvng forms such as rays (Raa sp.) have also been studed but, although generally less actve than dogfsh, they are more actve than Torpedo, whch may reman statonary n an aquarum for several days. Ths paper

86 G. M. HUGHES descrbes observatons on the respraton of ths fsh whch developed from ntal studes on the thornback ray (R. clavata). A partcular nterest n Torpedo from a physologcal pont of vew s ther response to hypoxa, as t was found that specmens reman nactve even at very low oxygen tensons and hence provde useful materal for nvestgatng the changes whch occur durng such extreme respratory stresses. MATERIALS AND METHODS Most studes were carred out at the Insttut de Bologe marne, of the Unversty of Bordeaux at Arcachon. Specmens of Torpedo marmorata Rsso, rangng n body weght from 500 g to 5 kg, were collected by fshermen from the Gulf of Gascony. They were kept n the sea-water crculaton of the laboratory and used at least week followng capture and were generally kept wthout feedng. Oxygen consumpton was measured usng three methods: () Contnuousflowresprometry. Fsh were kept n a ray-shaped resprometer and mantaned wth a constant flow of water of controlled temperature. The dfference n oxygen tenson between nlet and outlet water was montored contnuously by an oxygen electrode; samplng was manly of the outlet water wth occasonal samplng of nlet water by means of a tmng devce and samplng valve (Fg. a). (2) Closed chamber resprometry. By reducng the water flow to zero and closng the resprometer box, determnatons of the rate of fall n oxygen tenson enabled F Ol to be calculated (Fg. b). (3) Open chamber resprometry. In ths method fsh remaned n the same resprometer box but wth the ld removed so that gaseous exchange could take place at the ar-water nterface. The rate of declne n P 0} recorded by an electrode was lessened because of dffuson of oxygen nto the water from the ar. Calculatons takng ths nto account make t possble to estmate the oxygen consumpton of the fsh (Tamura & Morooka, 973). Torpedo was very sutable for ths method for, although t remaned statonary, the nspratory current at the spracles kept the water mxed (Fg. c). Followng anaesthesa wth MS 222, cannulaton was carred out n order to record pressure changes n the oro-branchal and parabranchal cavtes. The cannulae were brought out on the dorsal surface of the fsh. Samplng of water was also done usng these cannulae. Blood samples were obtaned followng cannulaton of a branch of the anteror mesenterc artery. The abdomnal cavty was opened by a medan ncson and the vscera dsplaced to one sde. A heparnzed cannula was nserted nto a small branch and passed nto the dorsal aorta. The cannula was brought out nsde a thcker tubng (pp. 20) nserted through the dorsal musculature. The abdomnal ncson was carefully sttched so as to prevent the entry of water. The fsh survved these operatons very well and there were very few losses for at least week followng the operaton. Blood samples could be taken wthout dsturbng the fsh n any way. Blood ph, P o and other parameters were measured on samples taken durng hypoxa and other experments usng a Radometer BMS 2 and oxygen electrode assembly. Oxygen dssocaton curves were determned usng a mxng method (Hughes, Palacos & Palomeque, 975). Pressure changes were recorded usng Sanborn 268 dfferental manometers and

Torpedo respraton (a) j- _ - Resprometer _j _- _* \ \ \ \ \ \ (c) - - Fg.. Oxygen consumpton measurements, (a) Dagram showng the apparatus used durng contnuous flow resprometry. Res., Upper reservor; P n, P out, samplng cannulae for nlet and outlet O a tenson; Stu., swtched valve; Cuv., cuvette contanng oxygen electrode. (6) Dagram of closed resprometer. (c) Dagram of open resprometer. Beckman LVDT couplers. They were dsplayed on a Tektronx 564 osclloscope and a Beckman 2-channel recorder. Smultaneous tape-recordngs of 4-channels enabled fuller analyss and recordngs to be made f requred at a later date. Analyss of the cardo/ventlatory couplng was carred out after recordng the data on punched-tape usng a Hewlett Packard 980A calculator (see Hughes & Adeney, 977). RESULTS (a) The mechansm of gll ventlaton As n Raa the man nspratory water current enters va the dorsal spracles, whch usually open synchronously durng each ventlaton cycle. The external openngs are much smaller than n many other rays and n some specmens there are perods when only a sngle spracle s functonal; such asymmetry was also observed n Raa clavata (Hughes, 960). The tme-course of pressure changes n the oro-branchal and parabranchal cavtes showed postve phases of relatvely short duraton. Expanson appears to be a largely passve process and occupes about four tmes the duraton of the actve contracton phase. Expanson of the cavtes s assocated wth a reducton n pressure and the tme course of changes n the parabranchal cavtes shows a more prolonged negatvty (Fg. 2). The waveforms n the oro-branchal cavty slghtly precede those n the parabranchal cavtes. As n other fshes the movement of water

G. M. HUGHES Seconds I + cm H 2 0 0 Dfferental pressure Parabranchal pressure Oro-branchal pressure E.C.G. Fg. 2. Smultaneous recordngs of pressure changes n the oro-branchal and parabranchal cavtes of Torpedo marmorata, together wth the dfferental pressure between them and the electrocardogram (E.C.C). across the glls s due to postve pressure pump and negatve sucton pump phases. The drect recordngs of the dfferental pressure show smlar waveforms to those derved for Raa clavata; the maxmum dfferental pressure beng very small (o--o-2 cm H 2 O). As n Raa, there does not seem to be a reversal phase. Durng hyperventlaton the ampltude of the pressure waveforms ncreases markedly (Fg. 9) and the dfferental pressure ampltude may be as great as 0-4 cm H 2 O. (b) Oxygen consumpton Most experments were carred out usng the contnuous flow method and results on dfferent szes offsh are plotted on log-log co-ordnates n Fg. 4. It s clear that the oxygen consumpton s almost drectly proportonal to body mass. Each of the ponts plotted on ths curve was derved from recordngs such as those shown n Fg. 3 (b). Followng an ntal equlbraton perod the oxygen consumpton shows some mnor fluctuaton but s generally farly steady (Fg. 3a). Mean values were taken at ths steady lower level of oxygen uptake, occasonal hgher values beng omtted from calculatons of the mean. It s apparent that ths restng level of oxygen consumpton may be mantaned over several days and s relatvely low when compared wth fgures n the lterature for other elasmobranchs (Table ). Comparsons of ths method wth the other two ndcated that closed chamber resprometry may produce hgher values, perhaps due to the accumulaton of carbon doxde and other products durng the experment. In general, however, the results compared qute well wth those of the contnuous flow method. The applcaton of the thrd method to fsh of ths knd for the frst tme s of specal nterest and gave qute

Torpedo respraton 80 60 ". (a) * * * * * «O 2 consumpton cm 3 /kg/h * «#.. 40 5 Temperature ( C) 0 5 - - 40 I 20 h 40 8 O 20 - m (b) I B o l l 8 0 2 4 6 Tme (h) o s " e o 0 ' 0 2 6 20 24 4 8 2 6 20 24 4 8 2 6 20 24 Tme (h) clavata. The dotted lne shows the mean level durng ths tme. (t>) Uxygen consumpton or three specmens of Torpedo marmorata over a perod of 2 3 days. Dfferent symbols refer to dfferent ndvdual fsh. smlar results. Ths method clearly has advantages over the closed resprometer n that there s less accumulaton of carbon doxde, but as n that method there wll be some accumulaton of other waste products whch would not be present wth the contnuous flow method. Some fgures obtaned for other speces usng ths method are gven n Table 2. (c) Relatonshps between the cardac and ventlatory rhythms Several studes of these rhythms n Torpedo were made many years ago (Schoenlen & Wllem, 894; Schoenlen, 895; Wllem, 92), and t was therefore of specal nterest to apply the method (Hughes, 972) developed for dogfsh to ths speces. Tape recordngs were made at regular ntervals over a perod of several days. The. J

9 G. M. HUGHES IOO 000 Body weght (g) 0 000 Fg. 4. B-logarthmc plot of restng oxygen consumpton of Torpedo marmorata aganst body mass; 95 % confdence lmts are gven for fsh of oo, 000 and 0000 g. Table. Comparson between oxygen consumpton of Torpedo and other elasmobranchs Temp. ( C) Body wt (g) (ml/kg, h) Reference Scylorhnus casscula S. cancula S. stellars Squalus suckley Torpedo torpedo T. marmorata 3 4 6 8 '3 4 6 6 6 4 260 49 22CO 660 IOCO 5000 40 448 803 3300 62-0 37 6o-o 380 55 38 32 49 54 2O-I 97 Wnberg, 956 Hughes & Umezawa, 968 a Baumgarten-Schumann & Pper, 968 Wnberg, 956 Prtchard, Florey & Martn, 958 Wnberg, 956 Ths paper Ths paper Ths paper

Torpedo respraton 9 Table 2. Comparson of values (mlo 2 /kg.h) obtaned for restng oxygen consumpton usng three dfferent methods n Torpedo marmorata, and other fshes Speces Torpedo marmorata "Serola qunqueradata "Epnephelus akaara 22-34 2O'O9 3-23 97 35 Contnuous flow Closed resprometry Open resprometry Wt(g) 26 803 3400 3300 43 Temp. ( C) 6 6 6 6 22'5 20-94 483 60 Wt(g) *Oplegnathus fascatus 26 S6 942 20-8 227 3700 27-2-27-8 From varous sources quoted, together wth other data, n Table 3-5 (Itazawa, 970). 26 62 445 Temp. ( Q 6-s 756 2859 3385 25-6 24-8 80 2-3 Wt(g) 26 803 3400 3300 7000 475 Temp. ( C) 4-5 85 65 65 recordngs were analysed as nterval hstograms (Fg. 0) and event correlograms plotted out on polar co-ordnates to show the phase angle and percentage couplng between the two rhythms. These plots ndcate the poston of the heart beat between successve oro-branchal waveforms (Fgs. 5, 6). In some cases the heart beat tends to occur n certan phases but at other tmes the rhythms seem much more ndependent and the couplng s very low. Under normoxc condtons the results were smlar to those found wth dogfsh,.e. ndvdual specmens showed dfferent degrees of couplng and phase angles whch also vared wth tme. Therefore t s not possble to make one generalzaton regardng the partcular relatonshp between the two rhythms whch s applcable at all tmes and for all ndvduals. Analyss of the recordngs suggested that greater couplng occurred at lower ventlatory frequences (Fg. 7) and also that the couplng was greater when the rato between two frequences was smpler,.e. towards whole number ratos (Fg. 8). (d) Response of the cardac and ventlatory rhythms to envronmental hypoxa The most obvous response of Torpedo to a lowerng of oxygen tenson of the nspred water was an ncrease n frequency of ventlaton whch was assocated wth a greater ampltude of these movements (Fg. 9). Analyss of nterval hstograms durng such hypoxa ndcated not only a reducton n the nterval between successve respratory cycles, but also a lessenng n the range of these ntervals (Fg. 0). As the rhythm became more rapd the cardac rhythm was not so regular both durng normoxa and hypoxa and showed relatvely small changes n the mean nterval. It s only n extremely prolonged hypoxa that a marked lengthenng of the nterval between the heart beats was found. Where such a bradycarda was not so well defned there was often greater rregularty n the cardac rhythm (Fg. n). Analyss of the cardo-ventlatory couplng durng hypoxa was carred out n detal and a comparson was made between the results of the automatc computer analyss and a more laborous cycle-by-cycle analyss of the same recordngs. Results of the latter are shown n Fg. 2, where the poston of each heart beat n a ventlatory 83 28-5

G. M. HUGHES Fg. 5. (a) Polar dagram showng varaton n couplng and phase relatonshps of an ndvdual specmen of Torpedo durng normoxa. (6) Polar dagram for sx specmens to show varaton n phase relatonshps ndcatng whch s most typcal for partcular ndvduals.

Torpedo respraton 93 60 50 40 a 30 20 0 90 80 270 Ventlatory phase angle 360 5 70 60 50 40 30 20 50 840 (b) \ / \-*- " 0 ' I. I 0.0 90 80 270 360 90 30 I 20 5 0 (c) S. " * -'. 0 90= 0 90 80 270 360 90 Ventlatory phase angle Fg. 6. Plots showng the percentage couplng and phase angle plotted n relaton to the tmecourse of an average oro-branchal pressure waveform. In (a) the data s plotted for sx specmens durng normoxa and n (b) for one specmen, manly durng normoxa and n (c) for four others. In (6) and (c) a tendency for the percentage couplng to fall as most heart beats occur later n the cycle can be recognzed, although the partcular relatonshp vares for the ndvdual case plotted. Dfferent symbols refer to dfferent ndvdual fsh. 70 r 50 x 0-24 h 24-48 h 48-72 h 30 0 5 20 25 30 Ventlatory frequency/mn Fg. 7. Plot of percentage couplng between the ventlatory and cardac rhythms aganst ventlatory frequency. All data s derved durng a 72 h perod from a sngle specmen (IX), manly under nonnoxc condtons and shows the fall n frequency durng acclmaton to the expermental condtons and the greater couplng at lower frequences. K.xn 73

94 V:H 2: G. M. HUGHES : :2 2: (b) : V: H 60 60 50 50 "o. o 30 U 20 M40 E 30 20 0 0 0 0-4 0-5 0-6 0-7 0-8 0-9 0-2 -3-4 5 0-5 0-6 0-7 0-8 0-9 0-2 Cardac frequency/ventlatory frequency Fg. 8. Relatonshp between percentage couplng and the rato of cardac to ventlatory frequency. In (a) data s plotted for four ndvduals whereas n (6) the data s obtaned from only a sngle ndvdual. A general tendency can be seen for the couplng to be greater when the rato between the frequences approaches a smple whole number. 4s 0 55 mm Ventlaton E.C.G. 63 mm mnlo 33 mm l\ 20 4. 34 mm 40 Fg. 9. Osclloscope recordngs of the oro-branchal pressure and electrocardogram durng a farly rapd hypoxa experment. The nspred Po 2 (mmhg) for each recordng s ndcated on the left, and the tme (n mn) followng the start of hypoxa s shown on the rght.

Torpedo respraton 95 Po, l52 2 _5 A- _«A_ 2 k 69 H 5 Tme (s) Fg. o. Torpedo. Interval hstograms for ntervals between successve ventlaton cycles (V) and heart beat ntervals (H) durng a rapd hypoxa experment wth recovery. Brackets ndcate pars of hstograms from smultaneous recordngs at the same Po 2 - Notce the decrease n nterval between ventlaton cycles and also reducton n the spread of the ntervals,.e. the hstogram covers a narrower range. The cardac rhythm remaned relatvely constant durng ths experment although a very slght lengthenng of nterval s recognzable. cycle s gven. The overall effect of hypoxa n ths partcular fsh was a sgnfcant reducton n couplng whch ncreased agan on restoraton of normal levels of nspred oxygen. Ths type of result was found n about 40 % of the fsh whereas the expected ncrease n couplng was found n only about 20%. It was only n later experments when Torpedo was subjected to prolonged perods at such low nspred oxygen tensons (less than 20 mmhg, for more than 4 h) that defnte evdence of bradycarda and some assocated ncrease n percentage couplng between the two rhythms was obtaned. Subsequent studes have demonstrated the presence of adaptatons at a bochemcal level durng such prolonged hypoxa (Hughes & Johnston, 977)- 4-2

G. M. HUGHES X EE 60-40- 20- p 25.20 5 W c ency /m 3 cr u.»«, t < < 2 3 N 2 on /.-- V \ > I If A / / '" V W W V / 4 /. \ \ H v \ 5 6 7 tn 2 off X I >! / ' / «/ / ft\ / ll I ^ / ^ / I- \\ / \ /Vv / V-- - J V V V /f 8 S 9 0 ^, - OB 2-5 - 20 - -5-0 r60-40 D. h20 a "c o 00-80- 60-40- 20- -= ( 0 20 40 60 80 00 Tme (mn) Fg. II. Analyss of a hypoxa experment showng changes n frequency of the heart (H) and ventlaton (V), and also changes n the percentage couplng (%C) between them. Ampltude (OB) of the oro-branchal pressure waveforms s also plotted. The ten perods ( o) durng whch couplng was computed are shown at the top of the graph and the results plotted at. Smlar analyses of the same tape recordngs over longer perods are shown at the bottom of the fgure. (e) Gll morphometry Glls were removed from the fsh ether mmedately after expermentaton or sometmes before any experments were carred out as the specmens were beng used by other vstors to the Insttute for studes on the electrc organs or sense organs. The glls were fxed n salt-water Boun and most measurements made later at Brstol usng methods that have been descrbed prevously (Hughes & Morgan, 973). The relatonshp between the man parameters of the glls and body weght (g) are plotted out n Fg. 3. The relatonshps for the man parameters are gven below: total gll area (mm 2 ) = 7-46W 0937, total flament length (mm) = 53-64W 0 ' 397, secondary lamellae/mm = 34-7^F~ ' 67. area of average secondary lamella (mm 2 ) = 0-00336 W 006.

, Torpedo respraton 97 t* K s N=44 '*,.*, '{ * * y -T 42 = max oo 86 '. ' 5 / 8 OO S? \ ' J 34 '.» 8 \.( J- *, t, «N ',,s \ 3C # / B V q. 4 '. L _ 9. t I X 40 m \ 40-30-. P Ol 5-54 92-0 [ 50-97J 3. 55-64 278-3 90 4. 64-37 2880 5-78 5. 37-28 mmhg 58-0 = 8 0-95= %C 20-0- q 40-30- 6. P Ol 28-00 275-4 4-64 7. 00-38 268-346 38-50 2500 4306 9. 50-54 24-6' 46-4 0. 54mmHg 23-5 = 6 58-2= %C 20-0- I I F Phases of ventlaton cycle Fg. 2. (a) Torpedo. Results of analyss of a sequence of ventlatory cycles durng a hypoxa experment. The poston of the electrocardogram wth respect of the ventlatory cycle n whch t falls s shown, together wth hstograms summarzng the overall dstrbuton wthn a cycle whch s dvded nto ten equal parts. The total number of heartbeats and number n the maxmum yg- cycle are gven for each analyss. (6) Detaled hstograms obtaned as ndcated n (a) together wth the phase angle and percentage couplng calculated for readngs durng whch the ambent P Oa was as ndcated. The numbers at the top of the four plots refer to the correspondng perods shown n Fg. n.

G. M. HUGHES loo IOOO Body weght (g) OOOO II 0 000 Sec. lamellae/mm 347 H'" 067 Ave. sec. lam. area =0-00336 W ' 706 0 000 0 000 Body weght (g) Fg. 3. B-logarthmc plots showng the relatonshp between (a) gll area and body mass and (b) the consttuent parameters of the gll area and body mass, for the same 22 specmens. In (6), upper crosses represent total flament length (mm), lower crosses number of secondary lamellae per mm, and flled crcles the area of the average secondary lamella (mm 8 ). Most ponts plotted refer to Torpedo marmorata, but fve of the measurements (enclosed n crcles) were made on large specmens of T. noblana. As these ponts seem to be qute close to the same general relatonshps they have been ncluded n the calculatons. Bars ndcatng 95 % confdence lmts for each of the regresson lnes are shown at body weghts of 00 and 000 and 0000 g.

Table 3. Summary of results of regresson analysts of gll area and ts component parameters, heart weght and restng oxygen consumpton aganst body mass (y = awb ) O-9399 08593 0-8926 0-9544 08554 95 % confdence lmts for fsh of 7-592 4-322 0-2622 0-05978 O-I543 0-0826 2-7425 0-899 Data gven prevously (Hughes, 976, 977) was based on only 0 specmens. 850-54 74466 II-3I33 0-3448 0594 O-3752 I-34I3 I-OI22 23-I458 37589 IOOOO g 8535467 504293-0 2757-73 8064-97 78487 6-9263 2-8:53 7873 5-04 96370 377-335 09-02 s 3 O 2 s- 3 Total gll area (mm 2 ) n 22 a 7-46 b 0-9368 Sa -6336 Sb 0-0634 Correlaton coeffcent W 0-960 00 g 3533-4 5694-378-65 2738-4 000 g 9 620-4 628606 Total flament length (mm) 22 53-64 0-3966 -306 0-0224 Secondary lamellae/mm (one sde of flament) Blateral area of an average secondary lamella (mm 2 ) Heart weght (g) 23 34-73 0-00336 0-0005 0-665 0-706 -038 0860-3497 0-440 0-05 0-0548 0-0484 Oxygen consumpton* (ml/h) 8 0-026 -0556 2-023 0-085

00 G. M. HUGHES 00 h 40-20 - 0 20 40 60 80 00 20 P Ol (mmhg) Fg. 4. Oxygen dssocaton curves for whole blood of T. marmarata at 5 and 20 C. The standard devatons for the P b0 values are shown. Statstcal data are shown n Table 3. It s evdent that the gll surfaces of a fsh weghng approxmately kg s about 80 mm 2 /g whch s very much less than that obtaned for most other marne fshes. It s also sgnfcantly less than correspondng measurements made for the thornback ray (R. clavata). Other parameters of the respratory and crculatory systems that were measured on ths specmen ncluded the heart weght whch was more or less drectly related to body weght: heart weght (g) = o-oobw 0 *. Although relatonshps for varous propertes of the blood and of the water blood/ barrer were nvestgated, no defnte relatonshp wth body weght was establshed. (/) Oxygen-carryng propertes of the blood As n prevous work on the thornback ray (Hughes & Wood, 974), most of the oxygen dssocaton curves were determned usng whole blood samples and a mxng method for determnng the P zo, P &0, P 70, P 80, etc. Because of changes whch occur n the propertes of blood followng ts removal from the fsh, only two determnatons were made on each of the blood samples, as ths can be carred out wthn half an hour of samplng. The oxygen dssocaton curves for whole blood, shown n Fg. 4, were obtaned by ths method usng blood equlbrated at two dfferent temperatures. The mean P 60 at normal physologcal condtons (5 C, ph 7-8) was 20-2 mmhg. The oxygen-carryng capacty s small (3-4 ml O 2 /oo ml blood) and consequently a far proporton of the oxygen carred n the blood s n soluton and ths has not been subtracted from the data plotted n Fg. 4. The Bohr factor A log P 60 /A ph, although

Torpedo respraton 0 Table 4. Relatonshp between oxygen consumpton (l^a) and surface area of the glls n fve speces of fsh Scylorhnus stdlars S. cancula Raa clavata Salmo gardner T. mamorata Restng (ml/kg, h) 38* 4*t 4 t 42-st 7-85 Gll area for kg fsh (mmvg) 35t 20Of 34t 97t 76 Baumgarten-Schumann & Pper (968). t Hughes, G. M. (977). X Randall, Holeton & Stevens (967). (ml/h.n 28-5 20 298-5 25-7 234-8 small, was qute defnte, havng a value of 0-32. The n vvo arteral blood oxygen tenson under restng normoxc condtons was usually about 70 mmhg and a ph of 7-82 at 5 C. More detaled studes on the characterstcs of blood wll be reported elsewhere. DISCUSSION Results of several types of nvestgaton on the respraton of Torpedo marmorata confrm the vew that t s a speces adapted to a sluggsh mode of lfe. Ths s n general agreement wth the lmted nformaton avalable about ts lfe habts but there s evdence that other speces of Torpedo (e.g. T. noblana) may be more actve. The oxygen consumpton of T. marmorata, measured by three dfferent methods, s lower than that of other elasmobranchs of comparable sze, as s the surface area of ts glls. Nevertheless, the rato between these two fgures (.e. the amount of oxygen transferred: unt area) s smlar to that obtaned for other speces (Table 4). The general rested behavour of Torpedo proved convenent n experments comparng three dfferent methods for measurng oxygen consumpton. The locomotory actvty was almost zero n all cases and the levels of O 2 uptake reman constant for long perods. The contnuous flow method s preferred but the other two methods are possble wth less equpment. The drecton of water flow, n at the spracles and out of the gll slts, created a current whch gave good mxng of the water. Ths was especally valuable n the use of Morooka's open resprometer method. Perhaps ths method could have wder applcaton, especally n the feld, where t mght be used to measure the oxygen uptake of sngle specmens of rare speces, e.g. Latmera, smply by followng changes n oxygen tenson of the water n a tank where t was kept alve. Ths method has the advantage over closed resprometry that the oxygen level does not fall to the same extent. Temperature control s mportant when usng oxygen electrodes, but can be dffcult n feld condtons. No attempt was made to nvestgate the effect of envronmental changes on oxygen uptake. It would be nterestng, for example, to know the effects of dscharge of the electrc organs upon restng oxygen consumpton. Ventlaton s acheved by mechansms essentally the same as those n Raa (Hughes, 960) but n ths case the spracular openngs are much smaller. The pressure changes have smlar relatonshps but the oro-branchal pump appears to be

02 G. M. HUGHES relatvely more mportant although the parabranchal cavtes clearly perform an mportant functon n suckng water through the glls. The low frequency and ampltude of the ventlatory movements are assocated wth a small dfferental pressure across the glls. The latter present a farly coarse seve of relatvely low resstance to water flow and consequently the ventlatory actvty must be relatvely economc n energy requrements. The regular and low frequency (0-5 mn) of the cardac rhythm probably has a smlar advantage. These features of the respratory system are smlar to those that are often found n other bottomlvng fshes. The respratory propertes of the blood reported here are also typcal of fsh havng relatvely sluggsh habts. The oxygen-carryng capacty s small and the hgh affnty blood has ts steepest part at oxygen tensons below 20 mmhg. Under normoxc condtons the arteral blood s more than 90 % saturated although the P Oa s only about 70 mmhg. Durng hypoxa the ntal response of Torpedo shows a defnte ncrease n both frequency and ampltude of the ventlatory movements. Changes n frequency are often qute marked (> 00%), whch contrasts wth ts relatve constancy n other elasmobranchs such as the dogfsh (Ogden, 945; Satchell, 96; Hughes & Umezawa, 968a; Pper, Baumgarten & Meyer, 970). Unlke other elasmobranchs (Satchell, 96; Pper et al. 970; Hughes, 973) bradycarda was ntally thought to be absent n Torpedo but of course t s more dffcult to recognse changes at such low frequences. Under more extreme hypoxa, both n the level of P o reducton and ts duraton, bradycarda becomes clearly recognzable below 20 mmhg. Under these more extreme condtons the ventlatory movements become reduced n frequency and ampltude and there follows a stage of ntermttent ventlaton before t ceases altogether. Durng the stage of perodc movements t s often observed that the fsh produces a 'spout' of water through the spracles whch s succeeded by several ventlatory movements of low ampltude. These observatons showng that the ventlatory rhythm s not so constant support the vew of Wllem (92) who dsagreed wth the statement of hs former collaborator (Schoenlen, 895) that the two rhythms moved together so that the pulse and breathng frequency always changed n the same drecton. However, the relatonshp between the two rhythms s by no means always a smple rato, whether ths be n the form of a synchronsm (.e. :) or the modfed form of synchronsm (2:,3:,4:) suggested by Wllem (92). However, the couplng tends to be hghest when the two rhythms approach such smple ratos. Under normoxc condtons some specmens show a relatvely hgh couplng, but under hypoxa when the ventlatory movements are accelerated and couplng tends to fall there s no correspondng change n the heart beat frequency. Couplng decreases durng hypoxa more often than t ncreases, a response opposte to that found n ranbow trout (Hughes, 973). Under the more prolonged and deeper hypoxa there s some ndcaton of an ncrease n couplng n some specmens. It s of nterest that Couvreur (902) was unable to fnd synchronsm between movements of the heart and respraton smlar to those reported for teleosts and cyclostomes. The nterpretaton of the sgnfcance of relatonshps between heart and ventlatory rhythms s more dffcult when the two rhythms have very dfferent frequences. An ncreased effectveness of gas transfer would be expected when the flow rates of water

Torpedo respraton 03 and blood concde. Such a relatonshp cannot possbly hold where the two frequences dffer very wdely. If the heart has a lower frequency than the ventlaton then t would clearly be advantageous for the perods of maxmum blood flow to concde wth perods of maxmum renewal of water at the gas exchange surface. The ventlatory cycles whch do not concde wth the heart rhythm would therefore be expected to have a lower effectveness. The analyses so far carred out are at a relatvely crude level concerned wth the bulk flows of the two meda whereas what s really mportant from a functonal pont of vew s the relatonshp between the flows of water and blood at the level of sngle secondary lamellae. Prelmnary bo-chemcal analyss ndcated an accumulaton of lactate, durng hypoxa and more detaled studes durng prolonged experments (Hughes & Johnston, 977, 978) have shown the presence of alternatve metabolc pathways n addton to the normal pathways of glycolyss. A sgnfcant ncrease of blood succnate may be the result of amno-acd fermentaton and has been recognzed as an adaptaton of a number of dvng vertebrates (Hochachka et al. 975). Anaerobc metabolsm has been shown to occur n a number of freshwater fshes (Blazka, 958; Johnston, 975) and fsh are known to lve n envronments where the oxygen content of the water s almost zero (Coulter, 967). Durng the course of the hypoxa experments t s unfortunate that no measurements were made of oxygen consumpton, especally durng recovery from the long hypoxa. From measurements durng some more extended closed and open resprometer experments there were ndcatons that the oxygen uptake falls when the oxygen n the water had been reduced below 50 mmhg. In vew of the adaptatons at a bochemcal level t s mportant to know whether ths fsh pays off ts 'oxygen debt' by oxdzng succnate and/or lactate. It would seem that Torpedo has a very hgh safety factor such that the oxygen tenson n the water can be lowered a great deal before any emergency mechansms are called nto acton. The ntal stages of regulaton would mantan respratory homeostass but once the capablty of ths mechansm has been surpassed the fsh seems to accept a more passve role untl, at even lower oxygen tensons, bochemcal mechansms come nto acton. An understandng of the overall adaptaton of Torpedo must take nto account all of these mechansms whch have so successfully become ntegrated that the fsh can reman apparently passve n the face of such extreme envronmental stresses whch n most other fsh would certanly ntate escape mechansms. T. marmorata generally mgrates nto the Bassn d'arcachon durng late summer when the water temperatures are farly hgh. It s often found at low tde n shallow mud pools near to the oyster beds. In these pools, as well as the bottom of the Bassn, Torpedo s largely covered by mud wth only the spracles vsble. If the electrc organs are used to paralyse ts prey then clearly t s a fsh whch s best able to survve by remanng statonary and quet. Thus t seems very lkely that at least some specmens would fnd themselves n envronments n whch the oxygen level was extremely low. Measurements would need to be made close to the spracular openngs f vald fgures were to be obtaned. Under other crcumstances t seems probable that certan speces of Torpedo can be more actve, presumably mgratng between dfferent parts of the sea bed.

04 G. M. HUGHES As always I am pleased to thank M. Bosseau and M. Cazaux and ther staff at Arcachon for generous and frendly help. I was also fortunate to enjoy the enthusastc research assstance of Rchard Adeney and Julan O'Nell durng most of ths work. Funds were provded by NERC and the Browne Fund of the Royal Socety. REFERENCES BAUMGARTEN-SCHUMANN, D. & PIIPER, J. (968). Gas exchange n the glls of restng unanaesthetzed dogfsh (Scylorhnus stellars). Respr. Physol. 5, 37-325. BEAMISH, F. W. H. (970). Oxygen consumpton of largemouth bass, Mcropterus salmodes, n relaton to swmmng speed and temperature. Can.J. Zool. 48, 22-228. BLAZKA, P. (958). The anaerobc metabolsm of fsh. Physol. Zool. 3, 7-28. BRETT, J. R. (964). The respratory metabolsm and swmmng performance of young sockeye salmon. J. Fsh Res. Bd Can. a, 83-226. BRETT, J. R. (972). The metabolc demand for oxygen n fsh, partcularly salmonds, and a comparson wth other vertebrates. Respr. Physol. 4, 5-70. COULTER, G. W. (967). Low apparent oxygen requrements of deep water fshes n Lake Tanganyka. Nature, Land. 25, 37-38. COUVREUR, M. E. (902). Sur le mdcansme respratore de la torplle. C. r. Sanc. Soc. Bol. 54, 252-253. HOCHACHKA, P..W., OWEN, T. G., ALLEN, J. F. & WHITTOW, G. C. (975). Multple end products of anaeroboss n dvng vertebrates. Comp. Bochem. Physol. 50 B, 7-22. HUGHES, G. M. (960). The mechansm of gll ventlaton n the dogfsh and skate. J. exp. Bol. 37,-27. HUGHES, G. M. (972). The relatonshp between cardac and respratory rhythms n the dogfsh, Scylorhnus cancula L..J. exp. Bol. 57, 45-434. HUGHES, G. M. (973). Respratory responses to hypoxa n fsh. Am. Zool. 3, 475-489. HUGHES, G. M. (976). On the respraton of Latmera chalumnae.j. Lnn. Soc. (Zool.) 59, 95 208. HUGHES, G. M. (977). Dmensons and the respraton of lower vertebrates. In Scale Effects n Anmal locomoton (ed. T. J. Pedley). New York, London: Academc Press. HUGHES, G. M. & ADENEY, R. J. (977). The effects of znc on the cardac and ventlatory rhythms of ranbow trout (Salmo gardner, Rchardson) and ther responses to envronmental hypoxa. Water Res. (In the Press.) HUGHES, G. M. & JOHNSTON, I. A. (977). Changes n blood chemstry of the electrc ray (Torpedo marmorata) durng hypoxa. J. Physol., Lond. 270, 69-70P. HUGHES, G. M. & JOHNSTON, I. A. (978). Some responses of the electrc ray (Torpedo marmorata) to low ambent oxygen tenson..?, exp. Bol. 73, 07-7. HUGHES, G. M. & KNIGHTS, B. (968). The effects of loadng the respratory pumps on the oxygen consumpton of Callonymus lyra.j. exp. Bol. 49, 603-65. HUGHES, G. M. & MORGAN, M. (973). The structure offsh glls n relaton to ther respratory functon. Bol. Rev. 48, 49-475. HUGHES, G. M., PALACIOS, L. & PALOMEQUE, J. (975). A comparson of some methods for determnng oxygen dssocaton curves of fsh blood. Revta esp. Fsol. 3, 83-90. HUGHES, G. M. & UMEZAWA, S.-I. (968 a). Oxygen consumpton and gll water flow n the dogfsh Scylorhnus cancula L. J. exp. Bol. 49, 557-564. HUGHES, G. M. & UMEZAWA, S.-I. (9686). On respraton n the dragonet, Callonymus lyra L.J. exp. Bol. 49, 565-582. HUGHES, G. M. & WOOD, S. C. (974). Respratory propertes of the blood of the thornback ray. Experenta 30, 67-68. ITAZAWA, Y. (970). Fsh Physology. Tokyo. (In Japanese.) JOHNSTON, I. A. (975). Anaerobc metabolsm n the carp (Carassus carassus L.). Comp. Bochem. Physol. 5 B, 235-24. OGDEN, E. (945). Respratory flow n Mustelus. Am. jf. Physol. 45, 34-39. PIIPER, J., BAUMGARTEN, D. & MEYER, M. (970). Effects of hypoxa upon respraton and crculaton n the dogfsh Scylorhnus stellars. Comp. Bochem. Physol. 36, 53-520. PRITCHARD, A. W., FLOREY, E. & MARTIN, A. W. (958). Relatonshp between metabolc rate and body sze n an elasmobranch (Squalus suckley) and n a teleost (Ophodon elongatus). J. mar. Res. 7, 403-4. RANDALL, D. J., HOLETON, G. F. & STEVENS, E. D. (967). The exchange of oxygen and carbon doxde across the glls of ranbow trout. J. exp. Bol. 46, 339-348. SATCHELL, G. H. (96). The response of the dogfsh to anoxa. J. exp. Bol. 38, 53-543. SCHOENLEIN, K. (895). Beobachtungen uber Blutkreslauf und Respraton be engen Fschen. Z. Bol. 32, 5-547.

Torpedo respraton 05 SCHOENLEIN, C. & WILLEM, V. (894). Observatons sur la crculaton du sang chez quelques possons. Bull. sc. FT. Belg. 26, 442-468. TAMURA, O. & MOROOKA, H. (973). The smple calculaton method of oxygen consumpton by aquatc anmals n water tank. I. Water volume constant. Bull. Fac. Fsh. Nagasak Unv. 35, 49-52. WEBB, P. W. (97). The swmmng energetcs of trout. II. Oxygen consumpton and swmmng effcency. J. exp. Bol. 55, 52-540. WILLEM, V. (92). Synchronsme des mouvements respratores et des pulsatons cardaques chez les possons. Acad. roy. Belg. Bull, des Set. 8, 508-533. WINBERG, G. G. (956). I. Rate of metabolsm and food requrements of fshes. II. New nformaton on metabolc rate n fshes. Fsh Res. BdCan. (translaton seres No. 94 and No. 362).